Intro/FAQ ⎜ Episode 1 ⎜ Episode 2 ⎜ Episode 3 ⎜ Episode 4 ⎜ Episode 5 ⎜ Episode 6 ⎜ Episode 7 ⎜ Episode 8 ⎜ Episode 9 ⎜ Episode 10 ⎜ Episode 11 ⎜ Episode 12 ⎜ Episode 13 ⎜ Episode 14 ⎜ Episode 15 ⎜ Episode 16 ⎜ Episode 17 ⎜ Episode 18 ⎜ Episode 19 ⎜ Episode 20 ⎜ Episode 21 ⎜ Episode 22 ⎜ Episode 23
[Editor’s Note: This is the twentieth in a series of 23 essays summarizing and evaluating Book of Mormon-related evidence from a Bayesian statistical perspective. See the FAQ at the end of the introductory episode for details on methodology.]
It seems unlikely that Book of Mormon place names would be so similar to places in the area around Palmyra.
In the early 1990s, a map presented by Vernal Holley purported to show place names in a broad area around Palmyra that were present in an altered fashion within the Book of Mormon. I demonstrate that, even with limited tools, it’s almost certain that one would find a false set of correspondences similar to Vernal’s given an area of similar size. Vernal’s map is essentially meaningless, and of zero use to those attempting to discern whether the Book of Mormon is authentic.
Evidence Score = 0
When last we left you, our ardent skeptic, you had just thrown the book down on the table, half in anger, half in fear that you’d gone mad. This book, whatever it was, has given you nothing that you could’ve expected—it felt as if the words had tapped into the occult, entrancing you and bending your own thoughts against you. But no, such a thing was impossible. There was no occult as surely as there was no God, and this book was no more mystical than the table on which it had unceremoniously landed. You’re tempted to leave it there—to prove that it has no power over you. But having gone this far, you’re determined to finish the book before it finishes you.
You sit back down, picking the book back up gingerly, trying—and failing—to find where you’d left off. After a few tries of thumbing through the book, you at last find yourself close to where you’d been just a chapter or so before. But on that page, a name manages to catch your eye. You hadn’t thought much of it your first time through, but now you can’t take your eye off it: “Teancum.”
As you repeat the word, it reminds you very much of Tecumseh, the famous Shawnee warrior that had unsuccessfully defied the U.S. government. Could the book’s author have based this Teancum, himself a fierce, courageous warrior, on the exploits of Tecumseh? It was possible, perhaps.
And that possibility is all you need. Before you continue reading, you decide to just take a look back through the rest of the book, to see if there are any names that ring a similar bell. You think that perhaps you’ll spend a minute or two satisfying your curiosity, but your curiosity can’t be tamed quite so easily. The hours pass as you comb back through the pages and, finding some loose sheets of paper on your shelves, you jot down a number of names that strike a familiar chord:
Teancum, Morianton, Ramah, Ripliancum, Ogath, Ephraim, Shurr, Cumorah, Onidah, Angola, Lehi, Shilom, Kishkumen, Jerusalem, Alma, Jacobugath
From there, you head back to your bookshelf and, in your excitement, begin hastily searching for your maps of the U.S. You soon find them and begin combing through them intently, searching for places that catch your eye. If this Smith was taking inspiration, perhaps he did so from the geography around him. You look first in the areas around western New York, where you knew Smith was when he dictated the book. Finding nothing, you expand your search somewhat to the whole of the state. And then to the surrounding states. Not yet satisfied, you go back to your stack of maps, pulling out the ones for Upper and Lower Canada. Yet more hours pass, and you find yourself smiling as you’re able to collect quite a few items of interest. You jot these down next to the original list:
Tecumseh, Moraviantown, Rama, Ripple Lake, Agathe, Ephram, Sherbrooke, Palmyra, Oneida, Angola, Lehigh, Shiloh, Kiskiminetas, Jerusalem, Alma, Jacobsburg
The correspondences weren’t exact, but they didn’t need to be. They were close enough for an enterprising Smith to refashion for his own purposes. The night has already gone dark and the air cold, but as your skin chills you have to work hard to force your attention away from your maps and your lists. You find yourself exulting as your numb hands begin to get the fire going. Had you at last found the weak link in Smith’s literary chain? You sit satisfied with your thoughts as you do your best to warm yourself by the growing flames. It sure seemed unlikely that one could find so many corresponding place names purely by chance.
Vernal Holley is far from the only person to assume that Joseph based the Book of Mormon off his own environment. But he’s the only one I’ve seen that’s decided to attach a map to that assumption. The map purports to show a number of places in the area around Palmyra that served as inspiration for places in the Book of Mormon, including the ones shown above. This map has been a rather soft target for apologetic voices over the years, enough that I hesitated to even address it. But since the map has been surprisingly persistent, I figured the dead horse was worth another kick.
Is Vernal’s map evidence of plagiarism? Are his correspondences at all unexpected, given the area in which Vernal was searching? Let’s see what Bayes can turn up.
Vernal Holley was an amateur historian and professional anti-Mormon, as well as the author of Book of Mormon Authorship, which theorizes the existence of a second manuscript written by Solomon Spaulding, which Joseph adapted to create the Book of Mormon. Based on the study of modern maps, Vernal created a hypothetical geography of the Book of Mormon, based on a set of 28 modern locations with varying degrees of similarity to names from the Book of Mormon. A version of this map can be found below:
A few things should be noted about this map and the places on it. First, some of these places are very small—barely blips on the map. They hardly seem like the kind of places that would have drawn Joseph’s attention. Second, despite purporting to align with the geography presented in the Book of Mormon, the map presented bears very little relation to the book’s internal geography, and contradicts that geography in a number of ways. Second, about half of the places identified by Vernal didn’t even exist at the time the Book of Mormon was written, greatly hindering Joseph’s ability to use them as inspiration. Third, the locations often bear only passing similarity to a minority of the 102 place-names in the book. You can see a summary of Vernal’s entire list below, including a rough measure of correspondence between the sets of names (the proportion of letters in the Book of Mormon names that are included in the locations identified by Holley):
|Holley Location||BofM Location||State/Province||Population||Number of Letters in BofM Location||Letters in Common|
|Conner||Comner||Unable to Locate||N/A||6||5|
|Saint Ephrem||Ephraim, Hill||Quebec||2567||7||5|
|Minonion||Minon||Unable to Locate||N/A||5||5|
|Proportion of Letters Represented:||86.5%|
We consider two relatively straightforward hypotheses.
Vernal’s correspondences arose on the basis of chance—Just as correspondences between two languages could arise by chance, so too could place-names similar to those in the Book of Mormon be found, particularly when many city-names both in the Book of Mormon and frontier America are biblically based. And the probability of a false correspondence would increase considerably as you looked at more and more names, in an increasingly broad search area.
Vernal’s correspondences are due to Joseph adapting local place names for use in the Book of Mormon—For this theory, a comparable case might be Tolkien’s world of Middle Earth, where Tolkien loosely based the locations and peoples of his world on those within his own surroundings in Europe. Though it’s likely hard to find obvious borrowing in Tolkien’s place names, as creative as he was, it’s at least conceivable that Joseph could have borrowed local place names as inspiration when writing the Book of Mormon.
PH—Prior Probability of Ancient Authorship—Our current estimate of the probability of an authentic Book of Mormon is 1 — 1.44 x 10-33. This, for those who haven’t been keeping track, is very close to 1. Here’s a look at the journey that’s gotten us here so far.
PA—Prior Probability of Modern Authorship—The remaining effectively-infinitesimal probability (p = 1.44 x 10-33) represents our current estimate for the likelihood of the Book of Mormon being a fraudulent modern production.
CH—Consequent Probability of Ancient Authorship—What we have here is a relatively straightforward problem, and a known process that should be fairly easy to replicate. We know that Vernal looked at modern maps covering a certain area around Palmyra, and that he looked for locations bearing similarity to Book of Mormon place names. Knowing that, we can do approximately the same thing. We can look at a different area of the U.S. covering about the same area (in square kilometers), and see whether we can find place names that show a comparable amount of similarity to that set of place names. If we can find as many places as Vernal did—ones that correspond to about the same degree—that should immediately invalidate Vernal’s model. If that same model could have arisen by chance just about anywhere he could have looked, then Vernal’s map is worth less than the paper it was written on.
Now, I’m not the first person to have tried to replicate this process. Shortly after the publication of Vernal’s book, Jeff Lindsay took a look at maps of Hawaii and was apparently able to find as many matches as did Vernal. But since I can’t seem to find the primary source for Lindsay’s attempt (and because I’m a glutton for punishment), I decided to try and see if I could replicate Vernal’s map on my own.
I’d feared that I’d be in for a few weeks of squinting at Google Maps, but luckily I found a bit more efficient way to tackle the problem. I started by taking a look at the Vernal’s full list, and trying to get a sense of the physical area that those cities cover. Though we could assume that Vernal would’ve been looking in a circular radius around Palmyra, we’ll be a little more conservative and assume that the area he was looking at was rectangular. That rectangle would stretch as far north and east as St. Ephrem, Quebec, as far west as Noah Lake, Michigan, and as far south as Alma, West Virginia. A little Google Maps magic helps us figure out that this rectangle measures 1,138 km by 761 km, for a total square area of 866,018 km2.
Obviously that rectangle features a substantial amount of water, so first we’ll need to remove that from our total area. If you add up the areas of Lake Huron (59,600 km2), Lake Erie (25,744 km2), Lake Ontario (18,960 km2), as well as that approximate rectangle of the Atlantic south of New England (which I measured using Google Maps at 208 km by 470 km for a total of 97,750 km2), and then subtract it from our total, that leaves a total land area of 663,954 km2.
The next task was to find a collection of U.S. states outside of that rectangle with the same approximate total land area. After a healthy application of guess-and-check methodology, I found four states that together total 663,427 km2. These states are Minnesota, Wisconsin, Missouri, and, of course, Florida. Now, rather than pore over maps of these areas, I went and found a handy public list of incorporated areas within those states, which included 3,611 place names. Fun Fact: if I’d paid $50 USD for the full list of U.S. place names, including unincorporated areas, I probably would’ve had between 4 and 5 times as many place names to work with.
Once I had that list of place names, I cross-referenced them with the list of 102 place names I put together for a previous episode, noting what appeared to me to be the closest match. Some of those were spot on, while for others I wasn’t able to find anything close to a good match. You can take a look at the full table in the Appendix if you’re curious.
Now, one thing I wanted to do was see if I could come up with some sort of measure of how close a particular match was. In Vernal’s case, there didn’t seem to be much rhyme or reason for seeing something as a particularly good match—sometimes the sounds matched, but sometimes they didn’t; sometimes there were extra words or syllables. In a way that makes a twisted sort of sense—not being in the mind of Joseph Smith it’s essentially impossible to know what internal criteria he would’ve potentially had for plagiarizing these place names, so there’s zero basis for picking one measure of correspondence over another.
But that’s not how we work around here. On the logic that even a crude measure was better than no measure, I went through Vernal’s examples and calculated the proportion of letters in the Book of Mormon place names that were accounted for in the modern place name. They didn’t have to match in terms of position (e.g., s as the third letter in both names)—as long as the letter was in the word at all it would count. As I suspected, Vernal’s names were (metaphorically) all over the map. There were 4 of Vernal’s 28 that were exact matches (e.g., Alma), and an additional 6 where all the letters in the Book of Mormon name that were represented (e.g., Morianton/Moriavontown). But there was some where as little as 60% of the letters were represented, such as Ripliancum/Ripple Lake. Overall, the average proportion of represented letters was 86.5%. So that was the bar that I wanted my set of correspondences to clear—for them together to have at least 86.5% of letters represented, and for all of them individually to be close to Vernal’s average (say, 80%). And of the 102 place names I identified in my set of states, 38 cleared that 80% bar.
I then went and did a more subjective pass on those 38 matches, pruning those that technically met the statistical criteria, but that didn’t quite pass the “smell test” of potential plagiarism (marked as “Prune” in the table in the appendix). In the end, I ended up with my own set of 28 correspondences, which together have an average of 92.6% letters represented. There were also an additional 5 matches that didn’t meet my benchmark of 80%, but that I still thought were fun, which I included as “honorable mentions.” Together, these comprise my soon-to-be world famous, totally-made-up Book of Mormon geography, complete with narrow neck, as shown below:
So, this is all a bit of fun, but what does that mean for our probability estimate? I considered for a bit trying to calculate the probability of a correspondence per square kilometer and simulating the distribution of matches that might be found in any comparable populated area. But at this point that wouldn’t be worth it. If I could do this with a relatively truncated list limited to incorporated areas, then doing this with the power of the most detailed maps would be child’s play. And with that full list it would be much easier to find places that were a plausible fit for the Book of Mormon’s internal geography (e.g., making sure my “Neck City” was actually close to a proposed narrow neck). I have to conclude that this exercise would be more than possible with any English-speaking area of equivalent size (and population density), particularly ones that utilize biblical names like Jacob or Ephraim. I don’t see any grounds for an estimate any lower than p = 1, meaning that this evidence is in no way, shape, or form unexpected.
CA—Consequent Probability of Modern Authorship—How likely would we be to find this evidence if the Book of Mormon was a modern production? Well, given that, again, we should be able to find this evidence almost literally anywhere, whether or not Joseph drew on those place names when writing the Book of Mormon, the probability shouldn’t be any lower than p = 1. But I’m really tempted to punish the critics for placing any credence in this evidence whatsoever. The theory that Joseph would’ve done something as silly as rip off local place names, while at the same time being creative enough to conjure—from whole cloth—a comprehensive and complex literary masterwork, is a theory deserving of ridicule. It’s not anything close to what we’d expect such an author to do—it’s a move just below what I’d expect from my nine-year old, to be honest. Of course, assigning a probability of anything less than 1 would make this evidence in favor of the Book of Mormon, and I can’t do that in good conscience. We’ll be assigning a consequent probability of p = 1.
The calculations this time around are going to be pretty dang easy.
PH = Prior Probability of the Hypothesis (our initial estimate of the likelihood of an authentic Book of Mormon, or p = 1 — 1.44 x 10-33)
CH = Consequent Probability of the Hypothesis (how likely we would be to observe these kind of place-name correspondences based on chance, or p = 1)
PA = Prior Probability of the Alternate Hypothesis (our initial estimate of the probability that the Book of Mormon was authored in the 19th century, or p = 1.44 x 10-33)
CA = Consequent Probability of the Alternate Hypothesis (how likely we would be to observe these kind of place-name correspondences whether or not the Book of Mormon is modern, or p = 1)
PostProb = Posterior Probability (our new estimate of the likelihood of an authentic Book of Mormon)
|PH = 1 — 1.44 x 10-33|
|PostProb =||PH * CH|
|(PH * CH) + (PA * CA)|
|PostProb =||(1 — 1.44 x 10-33 * 1)|
|((1 — 1.44 x 10-33) * 1) + (1.44 x 10-33 * 1)|
|PostProb =||1 — 1.44 x 10-33|
Lmag = Likelihood Magnitude (an estimate of the number of orders of magnitude that the probability will shift, due to the evidence)
Lmag = log10(CH/CA)
Lmag = log10(1/1)
Lmag = log10(1)
Lmag = 0
This evidence has no probative value when it comes to whether the Book of Mormon is authentic. May we never speak of it again.
Seriously, give this one a rest. It’s not worth your time.
Next Time, in Episode 21:
When next we meet, we’ll be looking at Book of Mormon names from another angle, discussing how likely we would be to have so many of those names show a meaningful connection to Hebrew and Egyptian, as well as how likely we’d be to see the variety in those names that we do if they were authored by a single person.
Questions, ideas, and ceramic dolls that stare blankly into your soul can be shipped care of BayesianBoM@gmail.com or submitted as comments below.
|#||BofM Place Name||Closest Match||State||Number of Letters||Letters in Common||Proportion of Letters Represented||Good Match|
|12||City by the Sea (unnamed)||N/A|
|18||East Sea||East Lake||Florida||4||4||100%||Yes|
|19||East Wilderness||East Williston||13||10||77%||No|
|27||Hagoth’s shipbuilding site||Hager City||6||3||50%||No|
|28||Head of the River Sidon||Sheldon||5||4||80%||Duplicate|
|33||Hill by Shilom||Shalimar||6||5||83%||Prune|
|36||Hill Riplah||Ripon||Wisconsin||6||3||50%||Honorable Mention|
|50||Land of First Inheritance||Inverness||11||6||55%||No|
|51||Land Southward||Southwest City||9||6||67%||No|
|65||Narrow Neck of Land||Neck City||Missouri||4||4||100%||Yes|
|66||Narrow Strip of Wilderness||Wildwood||10||4||40%||No|
|71||North Sea||North Bay||Wisconsin||5||5||100%||Yes|
|77||Plains of Agosh||Agency||5||2||40%||No|
|78||Plains of Heshlon||Hosford||7||3||43%||No|
|79||Plains of Nephihah||Neptune Beach||9||3||33%||No|
|87||South Sea||South Bay||Florida||5||5||100%||Yes|
|88||South Wilderness||Southwest City||15||8||53%||No|
|90||Tower of Sherrizah||Sheridon||9||6||67%||No|
|91||Valley of Alma||Alma||Wisconsin||4||4||100%||Yes|
|92||Valley of Gilgal||Gilman||6||4||67%||No|
|93||Valley of Shurr||Shullsburg||5||5||100%||Prune|
|94||Waters of Mormon||Morton||6||5||83%||Duplicate|
|95||Waters of Ripliancum||Rib Lake||Wisconsin||10||4||40%||Honorable Mention|
|96||Waters of Sebus||Sebastian||Florida||5||4||80%||Yes|
|97||West Sea||West Little River||4||4||100%||Prune|
|98||West Valley||West Vero||10||6||60%||No|
|100||Wilderness of Akish||Wilder||9||6||67%||No|
I hate to be simplistic, but let’s see if we can make some more horseplay on your 9-year-old child’s hypothesis.
So Joseph thinks to himself, “Let’s see, I need a name for a city. There’s Jacobsburg, Ohio on the map. Ah, Jacobugath. That’s really cool. No one will ever see it. Gosh that was so smart of me to do that. Except I’ve got to get back to that darned text. Why does everything have to be so hard? Now I’ve got to write a thousand word theme on my new city, Jacobugath, and their righteousness. Wait a minute, I can say that Jacobugath was named for a king named, aha, “JACOB!” Haha, it’s genius. And this city wasn’t righteous, it was evil, so it has to be destroyed. And why was it destroyed? Yes, because it had secret combinations! Not only will I destroy it, I’ll burn it. Yes, if I burn it, a hundred years from now, someone really smart will tie this in with volcanoes in meso-America! It’s brilliant! This is easier than I thought, thanks to my friendly map, I just pick a city and the whole text becomes so obvious it almost writes itself. Having a map makes everything so much easier…”
Hmmm, no wonder you didn’t want to go down that path. It’s pretty idiotic.
Given the likelhood that place names would have been most familiar orally in Joseph’s day rather than from written sources like in our time, it might be interesting to use a Soundex measure of name similarity. (Soundex is a comparison measure for proper names based principally on phonetic similarity.) But looking at the names, I doubt it would change your conclusion.
That’s a great idea, and one that I hope gets used on more productive problems than this one.